Automate Basics
New model · TechCrunch

Meta releases Muse Spark Open for local Mac workflows

By Automate Basics, written with AI from TechCrunch's original

· 2 min read

Mac Studio beside a camera and landscape footage in an editing room
Image: AI-generated illustration.

Meta has released Muse Spark Open, an open-weight AI model designed to run on Macs and handle coding and other desktop tasks. TechCrunch reports that running it locally requires unusually large amounts of memory.

What changed with Muse Spark Open

Meta has released Muse Spark Open, an open-weight AI model designed to run locally on Apple computers. According to TechCrunch, the model can work with different kinds of input and is intended for complex coding and tasks carried out across other software. Meta also makes it available through API platforms.

Meta says it paired the model with its Muse Code coding terminal for work on a Mastodon client, including fixes, features and a full build. The company also says it used the model with DaVinci Resolve to review and crop video footage. Those are examples Meta has reported, not a guarantee that the model will handle every project without supervision.

Who can use Muse Spark Open, and what does it cost?

Muse Spark Open is available now through API platforms, TechCrunch reports, although the article does not name those platforms or give prices. It also does not explain how someone can obtain the model for a local installation.

Local use is a much narrower option than the phrase “runs on Macs” might suggest. TechCrunch, citing Engadget, says a Mac needs at least 1 terabyte of RAM to run the model locally, and identifies an M5 Ultra Mac Studio as the best fit. For a team considering the model, the practical first question is whether it has access to suitable hardware or would need to use an API instead.

How do I try Muse Spark Open at work?

A coding team could start with a small, reviewable fix, while a video team could try selecting and cropping a short piece of footage. Those tasks reflect the uses Meta described and give a person a clear result to inspect. Before relying on a coding change, check that the build works and that the change does what was requested. For video, review the chosen footage and crop against the original.

TechCrunch does not provide setup instructions or name an API provider, so the article is not enough to plan a rollout. Teams should confirm how they would access the model, what their chosen route costs and whether their available Mac meets the local memory requirement.

Frequently asked questions

Can Muse Spark Open run on any Mac?

No. TechCrunch reports, citing Engadget, that running Muse Spark Open locally requires at least 1 terabyte of RAM. The article points to an M5 Ultra Mac Studio as the best fit, so a team should check its hardware before planning local use.

What work has Meta shown Muse Spark Open doing?

Meta says it used Muse Spark Open with its Muse Code terminal on coding work for a Mastodon client. It also reports using the model with DaVinci Resolve to review and crop footage. These examples show intended uses, but workers should still inspect the results.

Is there a price for Muse Spark Open?

TechCrunch says Meta is making Muse Spark Open available through API platforms, but the article does not give a price or name the platforms. It also does not explain how to obtain the model for local use, so those details need checking before a team budgets for it.

Written with AI from TechCrunch's original and published after automatic checks: every figure here appears in the original, and no sentence is copied from it. The picture is AI-generated. The original is the authority.

Source: Meta launches an open-weight AI model that runs on Macs, TechCrunch, 15 Oct 2026.